How Node Autocomplete helps you

When triggered from the Node Autocomplete Marker (purple sparkle), Node Autocomplete provides a list of nodes that may be good candidates for connecting to the port. This enables you to search a subset of viable nodes to find one that’s suitable for you, instead of having to browse through hundreds or even thousands of out-of-the-box nodes.

How to access Node Autocomplete

You can enable or disable Node Autocomplete from the Dynamo > Preferences > Features > Node Autocomplete menu.

Example of using node autocomplete auto wire feature

From these settings, you can select which UI you would like to use: - New UI - A new floating menu that allows for preview of the Node Autocomplete results. - Node Autocomplete Search UI - Traditional UI that looks like search within Dynamo. Also from these settings,you can select the default ranking method for Node Autocomplete results: Node Type Match or Recommended Nodes. For more information on these ranking methods, see the next section below.

When Node Autocomplete is enabled, you can access it by clicking on the Node Autocomplete Marker that appears on port hover. The Autocomplete flyout will appear next to the target port. To narrow down a large set of results, use the search field.



How the ranking methods work

Node Autocomplete provides two ranking method options: Node Type Match and Recommended Nodes. The ranking method determines the logic behind Node Autocomplete’s suggestions.

Recommended Nodes

The Recommended Nodes method suggests relevant nodes by using a recommendation engine powered by machine learning. Its suggestions are rooted in an analysis of node usage (node type, frequency of use) in thousands of sample graphs. More sophisticated than the Node Type Match method, Recommended Nodes is able to make more meaningful suggestions based on common and realistic graph practices.

The machine learning algorithm behind the Recommended Nodes method does not analyze user data. Instead, its recommendations are based on node use in a large set of sample graphs that the Dynamo team has selected.

Package support

Packages extend Dynamo functionality by allowing you to access and share custom nodes. In addition to the out-of-the-box set of Dynamo nodes, including nodes from host programs such as Revit, Node Autocomplete may recommend nodes from the following packages:

Node Autocomplete will only recommend nodes from packages that you currently have installed. Likewise, it will only recommend nodes that depend on host programs if you are running Dynamo from that host program. However, Node Autocomplete may sometimes recommend custom nodes from packages with Python scripts that depend on Revit library imports, even if you are running Dynamo Core (Sandbox). This is because there is currently no way for Node Autocomplete to identify which host program a Python-based custom node depends on.

Only the latest version of these packages is supported. If you have an older version of a package installed, any nodes from that package that have been deprecated in a newer version will not be included in the recommendations. To take full advantage of Node Autocomplete’s package support, make sure you have the latest versions of these packages installed.

Node Type Match

The Node Type Match method uses a ranking mechanism based on a node port’s type. This type, known as Object Type, tells the port what kind of data it can accept. Object Types include, among many others, string (text), integer (whole number), or object (such as a Revit wall). These types can be very specific, such as a Point.X node, which returns the X value of that point, or generalized, such as Geometry.Explode, which takes in any collection of geometry types and works on all of them..

   Nodes are sorted in this order:

  1. Nodes closest to the type (have the smallest type distance) that match the input/output port type.
  2. Nodes that are further away from the type (have a greater type distance)
  3. Group (such as Curve, List, or Math)
  4. Returned value (Output parameters)
  5. Name (alphabetically)
The result is an algorithmically ranked list of nodes that you can choose from via the Node Autocomplete dialog.

Which ranking method should I choose?

The Recommended Nodes method enables you to build your graph faster and with better results than Node Type Match. Beyond simple type matching, Recommended Nodes can provide more realistic and usable options, helping users of all experience levels make better node selections. The Recommended Nodes method will also continue to improve as the algorithm analyzes new sample graphs, making it an even more helpful design partner over time.

There may be times when the node you’re looking for doesn’t appear in the list of recommended nodes. In that case, try switching over to the Node Type Match ranking method to see a full list of nodes that match the triggering port’s expected data type. You can use Node Autocomplete’s search feature to narrow down the list of results.

For an additional consideration, the Recommended Nodes method generates network traffic, while Node Type Match does not. If you are experiencing network slowdowns while using the Recommended Nodes method, try switching to Node Type Match.

How to switch between ranking methods

You can change ranking methods at any time in the preferences menu.

Example of using node autocomplete feature

How the auto-layout mechanism works

Dynamo contains a feature called Cleanup Node Layout, which algorithmically cleans up your graph workspace to make all nodes and notes look cohesive through set distances and relationships.

To use Cleanup Node Layout, go to Edit in the Dynamo top menu and select Cleanup Node Layout. Alternatively, use the keyboard shortcut CTRL+L.

The node auto-layout mechanism with Node Autocomplete takes into consideration the entire node graph, paying particular attention to the local context, the parent node, and the margins between them before placing the new node. You may notice that your graph changes minimally with every new node placed through the Node Autocomplete feature. This means that you don’t have to manually move the placed node into a legible location, as all nodes will be visible and placed in a sensible location through this algorithm.


How the auto-wire mechanism works

When you place a new node on the workspace, Node Autocomplete will automatically attach the input/output port of the new node to the target output/input port where you accessed Node Autocomplete.

Example of using node autocomplete feature


How to explore Node Autocomplete Results

When you activate Node Autocomplete, you are able to explore suggestions by navigating with the arrows on the Node Autocomplete flyout, or arrow keys on your keyboard. (⬆️,⬇️,⬅️,➡️). To accept a result, simple press enter or select the checkbox on the flyout. To cancel the Node Autocomplete process, click the X on the flyout or press ESC on your keyboard.

Example of using node autocomplete feature


What nodes are hardcoded in and out of results

For nodes that do not match any of the result ranking criteria, such as some primitive type input/output ports, Node Autocomplete will show a fixed list of suggestions containing String, Number Slider, Integer Slider, Number, and Boolean.

Example of default list for node autocomplete feature